Proposed is a multi-view camera-based iterative calibration method for generation of a 3D volumetric model that performs calibration between cameras adjacent in a vertical direction for a plurality of frames, performs calibration while rotating with the results of viewpoints adjacent in the horizontal direction, and creates a virtual viewpoint between each camera pair to repeat calibration. Thus, images of various viewpoints are obtained using a plurality of low-cost commercial color-depth (RGB-D) cameras. By acquiring and performing the calibration of these images at various viewpoints, it is possible to increase the accuracy of calibration, and through this, it is possible to generate a high-quality real-life graphics volumetric model.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
2. The multi-view camera-based iterative calibration method of claim 1, wherein, in (a), the multi-view color-depth images include color-depth images of each viewpoint captured by at least four color-depth cameras constituting at least two horizontal layers, and numbers of the multi-view color-depth images in each layer are same.
This invention relates to a multi-view camera-based iterative calibration method for improving the accuracy of depth and color data captured by multiple color-depth cameras. The method addresses the challenge of aligning and calibrating multiple cameras to ensure consistent and accurate depth and color information across different viewpoints, which is critical for applications such as 3D reconstruction, augmented reality, and robotics. The method involves capturing multi-view color-depth images from at least four color-depth cameras arranged in at least two horizontal layers. Each layer contains the same number of cameras, ensuring uniform coverage and redundancy. The cameras in each layer capture synchronized color and depth data from different viewpoints, which are then processed iteratively to refine calibration parameters. The iterative process adjusts camera positions, orientations, and intrinsic parameters to minimize discrepancies between the captured depth and color data, improving overall accuracy. By using multiple cameras in a structured arrangement, the method enhances the robustness of the calibration process, reducing errors caused by occlusions, noise, or misalignments. The uniform distribution of cameras across layers ensures comprehensive coverage, allowing for more precise depth and color mapping. This approach is particularly useful in dynamic environments where real-time calibration is required. The iterative refinement ensures that the calibration remains accurate over time, even as environmental conditions or camera positions change.
3. The multi-view camera-based iterative calibration method of claim 1, wherein, in (c), the transformation parameter is optimized by calibrating point clouds of viewpoints adjacent in the vertical direction at each of all viewpoints in the horizontal direction, and a top-bottom matched point cloud is generated by matching and combining point clouds of viewpoints adjacent in the vertical direction with a coordinate system of one viewpoint among the viewpoints adjacent in the vertical direction.
This invention relates to a multi-view camera-based iterative calibration method for improving the accuracy of 3D point cloud data generated from multiple cameras. The problem addressed is the misalignment of point clouds captured from different viewpoints, which can lead to inaccuracies in 3D reconstruction or object detection tasks. The method involves capturing images from multiple cameras arranged in a grid-like structure, where cameras are positioned both horizontally and vertically. The key step is optimizing transformation parameters by iteratively calibrating point clouds from adjacent vertical viewpoints at each horizontal viewpoint. This ensures that point clouds from cameras in the same vertical column are aligned. Additionally, the method generates a top-bottom matched point cloud by matching and combining point clouds from adjacent vertical viewpoints, using the coordinate system of one of these viewpoints as a reference. This step ensures consistency between vertically adjacent point clouds, improving overall 3D reconstruction accuracy. The iterative process refines the calibration parameters to minimize misalignment errors across the entire camera array. The result is a more accurate and coherent 3D point cloud representation of the captured scene.
4. The multi-view camera-based iterative calibration method of claim 1, wherein, in (e), the point cloud of the virtual viewpoint is generated by combining some of top-bottom point clouds of two adjacent viewpoints of a corresponding viewpoint at each viewpoint in the horizontal direction when performing the virtual viewpoint calibration, and calibration is performed on each of the point cloud of the virtual viewpoint and the top-bottom point clouds of the two adjacent viewpoints of the virtual viewpoint.
This invention relates to a multi-view camera-based iterative calibration method for generating and calibrating point clouds from multiple camera viewpoints. The method addresses the challenge of accurately aligning and calibrating point clouds from different camera perspectives to create a coherent 3D representation. The process involves capturing images from multiple cameras positioned at different viewpoints, generating top-bottom point clouds for each viewpoint, and iteratively refining the calibration to improve accuracy. In the calibration step, a virtual viewpoint is created by combining top-bottom point clouds from two adjacent physical viewpoints. This virtual viewpoint's point cloud is then calibrated alongside the top-bottom point clouds of the adjacent viewpoints. The iterative process ensures that the point clouds from different viewpoints are accurately aligned, reducing discrepancies and enhancing the overall 3D reconstruction quality. The method is particularly useful in applications requiring high-precision 3D mapping, such as autonomous navigation, robotics, and augmented reality. By leveraging multiple camera perspectives and iterative refinement, the technique improves the reliability and accuracy of 3D point cloud data.
5. The multi-view camera-based iterative calibration method of claim 1, wherein, in (d), the top-bottom matched point clouds used for the round calibration are randomly extracted from a plurality of frames, and one of the top-bottom matched point clouds of consecutive frames of the respective multi-view color-depth images at each viewpoint is randomly extracted as a top-bottom matched point cloud of the corresponding viewpoint.
This invention relates to a multi-view camera-based iterative calibration method for improving the accuracy of camera calibration in systems using multiple cameras. The problem addressed is the challenge of achieving precise alignment and synchronization between multiple cameras, particularly in applications requiring high-accuracy 3D reconstruction or depth sensing, such as autonomous vehicles, robotics, or augmented reality. The method involves capturing multi-view color-depth images from multiple viewpoints and performing an iterative calibration process. During this process, top-bottom matched point clouds are generated by aligning depth information from overlapping regions of the images. These point clouds are then used to refine the calibration parameters iteratively. A key aspect of the invention is the random extraction of top-bottom matched point clouds from multiple frames to enhance calibration robustness. Specifically, for each viewpoint, one of the top-bottom matched point clouds from consecutive frames is randomly selected as the representative point cloud for that viewpoint. This random selection helps reduce bias and improves the accuracy of the calibration by ensuring diverse data is used in the iterative process. The method ensures that the calibration remains consistent across different frames and viewpoints, leading to more reliable 3D reconstructions and depth measurements.
7. The multi-view camera-based iterative calibration method of claim 1, wherein, in (c), (d) and (e), when performing the top-bottom calibration, the round calibration, and the virtual viewpoint calibration, respectively, one of the two viewpoints is set as a reference coordinate system, and the transformation parameter includes a rotation transformation matrix, a translation matrix, and a scaling factor for the reference coordinate system.
This invention relates to a multi-view camera calibration method that iteratively refines camera parameters to improve accuracy in 3D reconstruction or scene analysis. The method addresses challenges in aligning multiple camera views, such as misalignment due to physical movement, lens distortion, or environmental factors, which can degrade the precision of depth estimation or object tracking. The calibration process involves three stages: top-bottom calibration, round calibration, and virtual viewpoint calibration. In each stage, two camera viewpoints are selected, with one designated as a reference coordinate system. The method computes transformation parameters between the viewpoints, including a rotation matrix, translation matrix, and scaling factor, to align the coordinate systems. The rotation matrix adjusts angular discrepancies, the translation matrix corrects positional offsets, and the scaling factor compensates for size differences. These parameters are iteratively refined to minimize errors in the alignment process. The top-bottom calibration aligns vertically offset cameras, the round calibration adjusts cameras arranged in a circular or radial configuration, and the virtual viewpoint calibration synthesizes a new viewpoint from existing camera data. By iteratively applying these transformations, the method ensures consistent and accurate spatial relationships across multiple views, enhancing the reliability of 3D modeling, augmented reality, or robotic vision applications. The approach reduces manual intervention and improves automation in dynamic environments.
8. The multi-view camera-based iterative calibration method of claim 1, wherein, when performing the top-bottom calibration, the round calibration, and the virtual viewpoint calibration, the transformation parameter is optimized to minimize an error between an actual coordinate (Xref) of a point cloud of a reference coordinate system and a transformation coordinate (Xi′) by the transformation parameter.
This invention relates to a multi-view camera-based iterative calibration method for optimizing transformation parameters in a 3D imaging system. The method addresses the challenge of accurately aligning multiple camera views to reconstruct a precise 3D point cloud, which is critical for applications like augmented reality, robotics, and autonomous navigation. The calibration process involves three stages: top-bottom calibration, round calibration, and virtual viewpoint calibration. During each stage, the method iteratively adjusts transformation parameters to minimize the error between the actual coordinate (Xref) of a point cloud in a reference coordinate system and the transformed coordinate (Xi′) derived from the calibration process. This ensures that the reconstructed 3D model maintains high accuracy and consistency across different camera perspectives. The top-bottom calibration aligns vertically stacked cameras, correcting misalignments in elevation. The round calibration synchronizes cameras arranged in a circular or spherical configuration, ensuring seamless stitching of panoramic views. The virtual viewpoint calibration optimizes the transformation for synthetic viewpoints, enabling accurate 3D reconstruction from non-physical camera positions. By iteratively refining the transformation parameters, the method reduces discrepancies in the point cloud data, improving the overall fidelity of the 3D model. This approach enhances the reliability of multi-camera systems in dynamic environments where precise spatial alignment is essential.
10. The multi-view camera-based iterative calibration method of claim 9, wherein, in (g), when an error of the optimization function of all of the top-bottom calibration, the round calibration, and the virtual viewpoint calibration is within a predetermined range, and the variation leading to reduction of the error of the optimization function is less than a predetermined threshold variation when the repeating is additionally performed by a predetermined number of times, the ending is performed.
This invention relates to a multi-view camera system calibration method that iteratively refines camera parameters to improve accuracy. The method addresses the challenge of precisely aligning multiple cameras in a multi-view setup, which is critical for applications like 3D reconstruction, augmented reality, and autonomous navigation. Misalignment or calibration errors can lead to distorted or inaccurate depth perception and spatial mapping. The calibration process involves multiple stages: top-bottom calibration, round calibration, and virtual viewpoint calibration. Top-bottom calibration aligns cameras positioned vertically, while round calibration ensures consistency among cameras arranged in a circular or spherical configuration. Virtual viewpoint calibration adjusts parameters to optimize synthetic viewpoints derived from the multi-camera setup. Each stage uses an optimization function to minimize errors in camera alignment. The iterative process continues until two conditions are met: the error of the optimization function across all calibration stages falls within a predefined range, and the reduction in error from additional iterations is below a set threshold after a fixed number of repetitions. This ensures the calibration converges efficiently without unnecessary computation. The method dynamically adjusts parameters to achieve high-precision alignment, improving the accuracy of multi-view camera systems.
12. A non-transitory computer-readable recording medium having a program recorded thereon to perform the multi-view camera-based iterative calibration method for generation of the 3D volumetric model according to claim 1.
This invention relates to a computer-implemented method for generating a 3D volumetric model using multi-view camera calibration. The method addresses the challenge of accurately aligning multiple camera views to construct a precise 3D representation of an object or scene. The process involves capturing images from multiple cameras positioned at different angles, then iteratively refining the calibration parameters to minimize discrepancies between the captured views. This iterative approach ensures that the 3D model accurately reflects the spatial relationships between the cameras and the object being modeled. The method includes steps for initializing camera positions, estimating depth information from the images, and iteratively adjusting the calibration parameters based on the depth estimates. The goal is to achieve a high-fidelity 3D reconstruction by continuously refining the alignment of the camera views. The invention also includes a non-transitory computer-readable medium storing a program to execute this calibration method, enabling automated and efficient 3D model generation for applications such as computer vision, robotics, and augmented reality. The iterative refinement process improves accuracy by reducing errors in camera alignment and depth estimation, resulting in a more reliable 3D volumetric model.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
May 27, 2021
April 23, 2024
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.